Although detailed inference about the causal mechanisms of SRI use is not appropriate given the potential endogeneity concerns around some of the probit regressors, most of the variables included have the expected signs, signaling that the associations we expect indeed hold. Water availability, captured by dummies for upstream and midstream, as well as the dummy for the plot near the irrigation canal, are positive and significantly associated with SRI use, indicating that SRI tends to be applied when the supply of irrigation water is more reliable, and timely water management for alternate wetting and drying is much easier. The number of plots a household operates is negatively correlated with SRI use, presumably because the increased number of plots reduces the ability to closely manage each plot because this method requires greater cropping care and timely water management. The size of each plot is positively correlated with SRI use, implying that farmers appear more likely to experiment with a new method on larger plots, as found in previous studies (Barrett et al. 2004; Berkhout and Glover 2011; Noltze, Schwaltz, and Qaim 2012, 2013).